MS, University of Cincinnati, 2019, Engineering and Applied Science: Civil Engineering

The objectives of this thesis are to 1) Determine the extent to which commonly used building energy benchmarking metrics are correlated and 2) Establish if the added alternative metrics are valuable to the benchmarking scenario. If a high degree of correlation exists, then building performance under each metric would be similar and no new information is provided i.e. metrics are redundant. In this study, the results of dimension reduction analysis on commonly used building energy benchmarking metrics are examined. The thesis objective will be accomplished by applying a principal component analysis and factor analysis to a large database of buildings. The results of the analysis suggest that the most commonly used benchmarking metrics are correlated. Twelve selected metrics are reduced to three distinct factors, which are uncorrelated with each other. The first factor was related to the total energy consumption, the second factor was related to electricity, and the third factor was related to natural gas. Energy use intensity (EUI) and ENERGY STAR scores were captured together by the total energy consumption related factor. This implied that these vastly used metrics were redundant. The nine other metrics were loaded with the remaining two electricity and natural gas related factors and not the EUI and scores. It indicated that they are necessary for explaining the dataset and should be captured in the benchmarking discipline. New targets based on fuel types and end-uses are required for the complete building performance picture.

Committee: Amanda Webb Ph.D. (Committee Chair); Brock P. Glasgo PhD (Committee Member); Nabil Nassif (Committee Member) Subjects: Civil Engineering